Is inferential reasoning a distinctly human cognitive feature? Testing reasoning in cotton-top tamarins (Saguinus oedipus).

Abstract

Logical inference is often assumed a human-unique ability, although many species of apes and monkeys have shown some facility within a two-cup task in which one cup is baited, the primate is shown the cup which is empty (an exclusion cue), and subsequently chooses the other baited cup. In published reports, New World monkey species show a limited ability to choose successfully, often with half or more of the subjects tested not showing the ability with auditory cues or with exclusion cues. In this study, five cotton-top tamarins (Saguinus oedipus) were tested in a two-cup task with visual or auditory cues which revealed the presence or absence of bait, and in a second study, were tested with a four-cup array using a variety of walls to define the baiting space and a variety of visual cues including inclusion and exclusion. Tamarins demonstrated the ability to use either visual or auditory exclusion cues to find rewards in the two-cup study, although the visual cue required some exposure before accuracy was expressed. Experiment 2 revealed that two of three tamarins' first guesses to find rewards matched best a logic model. When they made errors, they typically chose cups adjacent to the cued location or made choices that seemed generated from avoiding empty cups. These results suggest that tamarins can deduce the location of food using reasoning, although the ability is only applied robustly to first guesses, while second guesses are motivated by approach/avoidance and proximity to cued locations.